While it’s a nice academic conversation, it’s hard to map the Aluminum Extrusions of some Deltas/Ingentis printers with the plywood of a Simpl printer, and waterjet cut pieces of a Makergear M2
It might be better to compare things like rigidity, or print speed, or ease of tuning, or even accessibility by the average DIY’er…I know from my standpoint, the parts I’ve inadvertently broken during my first build blows the value proposition WAY outta the water.
@Tim_Rastall @ThantiK
To be clear, i do think the dis advantages of owning a delta out weigh the advantages, i am not saying that i is a “better solution” i am saying that saying there is no advantage is wrong, there are advantages to delta printers, just their inherent problems are… well big problems.
You will have a better experience with a X-Y machine, but if you metric is $/cm3 and you care about print quality. I don’t think x-y wins.
In the $/cm3 metric, deltas are not inherently cheaper either. They just aren’t. There is literally no advantage to a delta machine other than “they look cool” (which isn’t really any kind of advantage at all!)…I surely haven’t been able to find any legitimate advantage, and I WORK FOR A COMPANY BUILDING THEM!
They’re great machines to look at! They DO look cool. But when we’re talking about mechanical, technical, cost advantages - They just simply do not exist.
Using the same linear rod lengths off of a delta, I can make a cartesian with 1.5x the build volume though yes, I may have to use 4 more linear bearings (a whopping $10 increase in cost, but we’re talking $/cm3 here, which means it’s better!)
@ThantiK how many linear rods would you need to build the printer of the same volume? yes the rods are the same length, but is the net length of all rods the same?
As i said, i don’t think that they are over all a better design. i think that there much much better choices, but that is just me.
@Camerin_hahn yes, with equal length rods, the same number of them. In fact, I did it with my rostock mini. You can do the calculations yourself - how much linear rod it takes to build a delta of X/Y/Z, and then how much of that rod could build what dimension machine in Cartesian format. Cartesian will always win unless it’s a stupid design, like a moving X/Y bed.
@Camerin_hahn I’ve never built a Delta and I’m prepared to concede they trump Cartesian machines in several areas but I don’t think print quality is one, without even getting started on the position dependant variable resolution of deltas, the best quality I’ve seen from desktop FDM is the UP plus and anecdotally, people are saying the Zortrax m200 is doing even better.
The length of the effector arms (which determines the build area in the X/Y plane) is effectively subtracted from the Z travel, so the effective minimum height of the machines (assuming zero carriage length and that the carriages can move all the way to the end of the machine’s dimensions, which they never can), the minimum height of a delta is the Z travel plus the X/Y radius, and nearly all of the X/Y space can be used (again, ideally; in practice, the size of the effector plus the size of the carriages and linear rails on each side mean you end up with closer to half the X/Y space the machine takes up, less if you consider the desk space required for the machine as a square, in which case (1 - PI/4) * the machine’s radius is effectively wasted space).
For a cartesian bot with an overhead X/Y gantry, the theoretical maximum usable space on any axis is the length of the machine on that axis minus the corresponding dimension of the toolhead.
@ThantiK I was forgetting that the print “radius” as i cant thing of the right term, needs to be added to the top of the delta at least 1 once (more in practice) for the movement to work out, that means you have 3 times the x-y there and what 1 times larger z height… I was wrong there
then compound that with the movement speed being slow in the middle and jerky and fast at the edges. you have lots of problems.
out of curiosity do you feel the same way about the Simpson. it is a delta, it has most of the flaws that you would have in the tower structure, but it does have the benefit that you are not doubling back on the you tower, and the detraction that you are increasing the foot print (also looks cooler then a delta tower)
The delta kinematics for a simpson are a little different. I haven’t done the math as to whether this works out to a net positive or negative. The latter seems more likely, but I assume the odd way that the motors are coupled to the axis drive is designed to mitigate this.
I feel a lot more positive about the Simpson simply because of its RepRap aspects. The simpson doesn’t inherently use less parts, but it does use parts that are much easier and cheaper to obtain. As far as delta calculations go, I’d be willing to bet that it has the same drawbacks as traditional delta machines. However, mechanical advantage can be added in order to increase the resolution and mitigate some of these effects. The Simpson has a huge gearing ratio for that reason. You could largely also offset the effects of a standard delta by using screws instead of belts, which might actually work because since the motion of the end effector is the sum of all of the motions on each arm you aren’t severely limited in speed by gearing down.
tl;dr delta resolution problems can be mitigated by mechanical gearing
You can mitigate the loss of resolution, but then you run into step rate limit issues on the other end.
The Simpson uses parts that are cheaper and easier to obtain if you already have a working printer, but are next to impossible to obtain otherwise. And forget mass production.
@Whosa_whatsis I think that a Simpson variant could be mass produced, but you would need to do allot of work. you would need to injection mold the plastic parts for sure as print time is way to long for making 1 let alone 100 of them.
@ThantiK @Tim_Rastall I do like the anti-anti-Cartesian group idea. it sound just silly enough to be fun
@Camerin_hahn I’d guess the wally would be a better mass production candidate, the componentry is much less complex and would be easier to produce from an injection moulding perspective, (as far as I can make out).
For every great delta printer, there is always be a better cartesian printer. 